Difference between revisions of "Google Code In/2016/TODO"

(track what we need for gci)
 
(mention opencl major task set)
 
Line 8: Line 8:
 
# '''get the virtual disk image updated/rebuilt or get a docker image set up'''<br>this is so participants can be given a fully prepared environment with the fewest instructions possible.<br>
 
# '''get the virtual disk image updated/rebuilt or get a docker image set up'''<br>this is so participants can be given a fully prepared environment with the fewest instructions possible.<br>
 
# '''prepare major task sets'''<br>we have a dozen or so "big jobs" but we have to do the first one for them, to establish the desired pattern.  lets brainstorm them here:
 
# '''prepare major task sets'''<br>we have a dozen or so "big jobs" but we have to do the first one for them, to establish the desired pattern.  lets brainstorm them here:
 +
## consolidate simple argc/argv functions together (LIBGED refactoring)
 +
## convert C functions to OpenCL (performance pipeline work)
 
## turn LIBGED functions into plugins (src/libged/zoom is close, needs to be finalized)
 
## turn LIBGED functions into plugins (src/libged/zoom is close, needs to be finalized)
 
## debug and fix a single unit test failure (NURBS Boolean)
 
## debug and fix a single unit test failure (NURBS Boolean)
## consolidate simple argc/argv functions together (LIBGED refactoring)
 
 
# '''write up mentor rules and guidelines'''<br>need to make sure we're all on the same page, reviewing consistently, the the same quality expectations.
 
# '''write up mentor rules and guidelines'''<br>need to make sure we're all on the same page, reviewing consistently, the the same quality expectations.

Latest revision as of 20:55, 15 November 2016

  1. generate task ideas
    we have a lot from previous years, but need better focus and to make sure they're all appropriate. the start of this is to simply brainstorm ideas in a spreadsheet. emphasis on easier to understand titles (NUA - no undefined acronyms).
  2. get a repo space set up for integrating GCI work
    this is either an svn branch or a git clone/bridge. whatever the solution, the fewer steps for participants to actually getting committable work, the better.
  3. get our source code set up
    1. run astyle on the entire tree, eliminate mixed tab+space indentation
    2. update docs, snippets, etc, to new style
    3. update cmake build to test/report style errors during compilation
  4. set up a design gallery
    we need to be able to showcase completed GCI work somewhere. this is so others can make theirs better than those already accepted, and it makes it easier to detect blatant copying. probably can just use yash's gallery from GCI 2014, or build something
  5. get the virtual disk image updated/rebuilt or get a docker image set up
    this is so participants can be given a fully prepared environment with the fewest instructions possible.
  6. prepare major task sets
    we have a dozen or so "big jobs" but we have to do the first one for them, to establish the desired pattern. lets brainstorm them here:
    1. consolidate simple argc/argv functions together (LIBGED refactoring)
    2. convert C functions to OpenCL (performance pipeline work)
    3. turn LIBGED functions into plugins (src/libged/zoom is close, needs to be finalized)
    4. debug and fix a single unit test failure (NURBS Boolean)
  7. write up mentor rules and guidelines
    need to make sure we're all on the same page, reviewing consistently, the the same quality expectations.